e7959c1935
We have two drivers for a 100%-identical peripheral right now, mostly because we couldn't come up with a good common name for it back when we checked it in. That seems like a pretty silly reason in the long run. Both Tegra and Rockchip SoCs contain UARTs that use the common 8250 register interface (at least for the very basic byte-per-byte transmit and receive parts we care about), memory-mapped with a 32-bit register stride. This patch combines them to a single 8250_mmio32 driver (which also fixes a problem when booting Rockchip without serial enabled, since that driver forgot to check for serial initialization when registering its console drivers). The register accesses are done using readl/writel (as Rockchip did before), since the registers are documented as 32-bit length (with top 24 bits RAZ/WI), although the Tegra SoC doesn't enforce APB accesses to have the full word length. Also fixed checkpatch stuff. A day may come when we can also merge this driver into the (completely different, with more complicated features and #ifdefs) 8250 driver for x86 (which has MMIO support for 8-bit register stride only), both here and in coreboot. But it is not this day. This day I just want to get rid of a 99% identical file without expending too much effort. BUG=None TEST=Booted on Veyron_Pinky and Nyan_Blaze with and without serial enabled, both worked fine (although Veyron has another kernel issue). Change-Id: Ib84d00f52ff2c48398c75f77f6a245e658ffdeb9 Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/225102 Reviewed-by: Aaron Durbin <adurbin@chromium.org> |
||
|---|---|---|
| .. | ||
| serial | ||
| storage | ||
| timer | ||
| usb | ||
| video | ||
| cbmem_console.c | ||
| hid.c | ||
| keyboard.c | ||
| Makefile.inc | ||
| nvram.c | ||
| options.c | ||
| pci.c | ||
| speaker.c | ||